le même processus pour tous - Université de Bourgogne
le même processus pour tous - Université de Bourgogne
le même processus pour tous - Université de Bourgogne
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
702 M. Lelièvre et al. / Food Quality and Preference 19 (2008) 697–703<br />
4.1. Are trained and untrained assessors comparab<strong>le</strong>?<br />
To address this question, we compared trained and untrained<br />
assessors <strong>de</strong>scriptions. In the condition without list, we found that<br />
the <strong>de</strong>scriptions of the groups of beers were rather different for<br />
both groups of assessors. This result does not replicate Cartier<br />
et al. (2006) study which found that the <strong>de</strong>scriptions of groups of<br />
breakfast cereals were almost similar between trained and untrained<br />
assessors. We observed that there were many more terms<br />
quoted by untrained assessors (54 terms) than by trained assessors<br />
(35 terms). But when se<strong>le</strong>cting only terms with a geometric mean<br />
above 20%, only three terms for untrained assessors and eight terms<br />
for trained assessors were kept. This result ref<strong>le</strong>cts the lack of consensus<br />
in both the choice of the terms and in perceived intensity,<br />
especially for untrained assessors. The greater lack of agreement<br />
among untrained assessors in comparison to the trained assessors<br />
is not very surprising. In<strong>de</strong>ed, training involves the <strong>de</strong>velopment<br />
of a common <strong>le</strong>xicon with standard physical references allowing<br />
an alignment and a standardization of the sensory concepts of the<br />
panelists (Ishii & O’Mahony, 1990). The importance of training in<br />
reaching a consensus is illustrated by the fact that seven out of<br />
the eight terms of trained assessors were attributes belonging to<br />
the profi<strong>le</strong> list of attributes used for their training. For examp<strong>le</strong>, a<br />
trained assessor <strong>de</strong>scribed the three Leffe beers in this way: ‘‘very<br />
sweet, very alcohol, medium hop, medium bitter,” whereas an untrained<br />
assessor <strong>de</strong>scribed these same beers with: ‘‘medium exotic<br />
feel, medium spicy sensation, medium grapping taste (goût prenant).”<br />
This difference between the <strong>de</strong>scriptors used by trained<br />
and untrained assessors can be explained by the training of trained<br />
assessors which allows them to possess a specific and precise<br />
vocabulary. Finally we found that trained and untrained assessors<br />
used the four intensity words differently. Contrary to untrained<br />
assessors who used the three expressions ‘‘a litt<strong>le</strong>,” ‘‘medium,”<br />
and ‘‘very” in the same way, we observed that trained assessors<br />
used ‘‘very” twice as often as ‘‘a litt<strong>le</strong>.” We also noticed that untrained<br />
assessors used the word ‘‘not” frequently, whi<strong>le</strong> trained<br />
assessors hardly used it. So it seems that trained assessors tend to<br />
<strong>de</strong>scribe their groups of beers with distinctive characteristics (i.e.,<br />
characteristics with a high intensity) whereas untrained assessors<br />
do not use particular characteristics to <strong>de</strong>scribe their groups of<br />
beers. These observations highlight the interest of using intensity<br />
scores to quantify attributes. These quantitative words bring additional<br />
information to the <strong>de</strong>scriptions and we think that it is important<br />
to impose their use to the assessors.<br />
The comparison between trained and untrained assessors’<br />
<strong>de</strong>scriptions confirmed the conclusions of several authors that<br />
trained assessors used more specific terms, especially terms<br />
<strong>le</strong>arned during training (Chol<strong>le</strong>t & Va<strong>le</strong>ntin, 2001; Chol<strong>le</strong>t et al.,<br />
2005; Clapperton & Piggott, 1979). We expected this high specificity<br />
of trained assessors’ vocabulary to <strong>le</strong>ad to a better matching<br />
performance than that of untrained assessors. Yet, contrary to previous<br />
work (Gawel, 1997; Law<strong>le</strong>ss, 1984; Solomon, 1990) we did<br />
not find any difference in matching performance between the<br />
two groups of assessors. Both trained and untrained assessors were<br />
above chance <strong>le</strong>vel but their performance <strong>le</strong>vels were not very high<br />
(54.7% of correct matches for trained assessors and 50.9% for untrained<br />
ones). The overall low performance of trained assessors,<br />
however, might be due to the high inter-individual variability. In<strong>de</strong>ed,<br />
whi<strong>le</strong> three trained assessors performed perfectly, two others<br />
were below chance <strong>le</strong>vel. A plausib<strong>le</strong> explanation for this<br />
high variability is the difference in years of training of the panelists.<br />
In<strong>de</strong>ed, the panelists with 100% of correct matches were<br />
among the panelists who had the longest training. Yet correlation<br />
coefficient computed between the percentage of correct matches<br />
and the years of training shows that it is not the only explanation<br />
(r = .61, r 2 = 0.37, p < .05). The fact that some trained assessors with<br />
four or five years of training succee<strong>de</strong>d in the matching task<br />
whereas others had poor results may suggest that some trained<br />
assessors are better than others to generalize their know<strong>le</strong>dge to<br />
a new task. It has been already showed that trained assessors were<br />
not ab<strong>le</strong> to generalize their perceptual know<strong>le</strong>dge to new beers<br />
(Chol<strong>le</strong>t et al., 2005). The same prob<strong>le</strong>m could exist with new tasks<br />
and this might be related to the duration of training.<br />
4.2. Is providing a list helpful?<br />
We found that the <strong>de</strong>scriptions of the beers were different<br />
when assessors had a list of terms and when they did not have<br />
such a list, especially for untrained assessors. For untrained assessors,<br />
we observed a larger number of <strong>de</strong>scriptors with a geometric<br />
mean above 20% with the list than without the list. This suggests<br />
that having a list of terms can be helpful for untrained assessors.<br />
But a <strong>de</strong>eper look at the <strong>de</strong>scriptions with the list shows that,<br />
for examp<strong>le</strong>, untrained assessors used hop and malt to <strong>de</strong>scribe almost<br />
all the beers. It is probab<strong>le</strong> that the list given to untrained<br />
assessors influenced their <strong>de</strong>scriptions. The untrained assessors<br />
probably knew that hop and malt are terms associated with the<br />
brewing process and so they used it but without knowing exactly<br />
what these terms mean. We assume that the <strong>de</strong>scriptions containing<br />
these words hop and malt did not allow them correct matches.<br />
For trained assessors, the number of <strong>de</strong>scriptors with a geometric<br />
mean above 20% was quite similar between the two conditions.<br />
Moreover, the results of the matching task were not better with<br />
the list than without the list for both trained and untrained<br />
assessors.<br />
The efficiency of the list in this study can be put in perspective<br />
with the results of Hughson and Boakes (2002). In this<br />
study, assessors had to <strong>de</strong>scribe five white wines according to<br />
three conditions: without any list of terms, with a long list of<br />
terms (125 terms) and with five short lists of terms (14 terms<br />
in each list corresponding to each wine). Then, they had to<br />
match their own <strong>de</strong>scriptions to the wines. Matching performance<br />
was better in the short-list condition (40% of correct<br />
matches) than in the long-list condition (27% of correct matches)<br />
and in the control condition without any list (16% of correct<br />
matches). Moreover, only results in the short-list condition were<br />
above chance. So we can won<strong>de</strong>r why our list did not help assessors<br />
to improve their scores of matching too. One reason could<br />
be that our list of terms was too long (44 terms) compared to<br />
the one of Hughson and Boakes (14 terms) to help assessors to<br />
effectively <strong>de</strong>scribe the beers. In the case of trained assessors,<br />
another reason could be that the terms provi<strong>de</strong>d were different<br />
from the terms used in training. This hypothesis is supported<br />
by the fact that trained assessors <strong>de</strong>scribed ChtiBL as butter in<br />
the condition without the list but did not in the condition with<br />
the list. Interestingly, in Meilgaard’s list, butter is replaced by<br />
diacetyl, which is associated with the butter flavor and so trained<br />
assessors did not seem to know the term diacetyl. This remark<br />
highlights the importance of using a common <strong>de</strong>scriptive vocabulary.<br />
Some authors such as Rainey (1986), Civil<strong>le</strong> and Law<strong>le</strong>ss<br />
(1986) or Stampanoni (1994) indicated that for sensory profi<strong>le</strong>s,<br />
the use of a common terminology based on references reduced<br />
the time for training and improved the agreement between the<br />
assessors. In our case, the use of a terminology without associated<br />
reference did not help assessors to <strong>de</strong>scribe the beers.<br />
Finally, the fact that the list of terms did not help the assessors<br />
could be due to the use of a previously published list which was<br />
not exactly adapted to our products. In the study of Hughson<br />
and Boakes (2002), the short lists provi<strong>de</strong>d to the assessors contained<br />
terms which correspon<strong>de</strong>d exactly to the wines to be<br />
<strong>de</strong>scribed.